Consider switching from cutadapt to pTrimmer

[donkirkby]

With the changes for SARS-CoV-2 in #549, we may need to trim many more primer sequences. Look at pTrimmer and its paper to see if it's more efficient than cutadapt.

• Trim ARTIC primers with cutadapt after trimming adapters.
• Compare pTrimmer results with cutadapt results.
• Add HCV primers.
• Fix microtest examples to work with new primer trimming.
• Time cutadapt and pTrimmer calls for FASTQ files of sizes 20MB, 100MB, and 500MB, as well as SARS-CoV-2, HCV, and HIV.
• Add options to skip different trimming steps: phiX, adapters, and primers.

[donkirkby]

The primer sequences to use for SARS-CoV-2 are in the artic project, and SRR11314339-SARS_S12 is a (poor) example of ARTIC/PrimalSeq data. (Since it was processed with Nextera XT it is likely that some read pairs will contain zero or one primer, rather than 2.)

[donkirkby]

Clarification from slack discussion:

• Adapters have different rules from primers.
• Primers only appear at the ends of reads, adapters can appear in the middle of a read if you read off the end.
• Trim adapters first, then primers.
• Primer-dimer scenario could look like [LEFT]-[RC RIGHT]-[ADAPTER]-[random garbage]

Check the documentation of adapter types.

[donkirkby]

Time in minutes for cutadapt and other trimming steps:

sample	read count	censor	adapters	primers
S8CT25MIXATOH-NT-Unknown_S29	108,287	1	0	2
P90786A-HCV_S86	475,577	3	1	8
S3CT27-IP-Unknown_S8	482,083	3	0	9
HIV3428H2-E21-HIV_S39	544,536	4	1	10
POS1TO1K-IP-Unknown_S2	1,632,059	20	2	30
POS1TO1KMIX-IP-Unknown_S4	1,751,820	21	2	32
POS1TO20K-IP-Unknown_S1	2,228,806	27	3	40
POS1TO20KMIX-IP-Unknown_S3	4,376,862	52	5	78